The Agent Skills Directory

[Prompt Injection] (SAFE): The instructions are focused on Rails testing procedures and do not contain any attempts to bypass safety filters, ignore previous instructions, or extract system prompts.
[Data Exposure & Exfiltration] (SAFE): No commands access sensitive file paths (e.g., ~/.ssh, .env) or initiate network connections to external domains.
[Obfuscation] (SAFE): The skill is written in clear markdown with no Base64, zero-width characters, or encoded strings.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill does not download external scripts or install third-party packages. It relies on local bin/rails commands which are standard for the Rails environment.
[Indirect Prompt Injection] (LOW): The skill instructs the agent to read files in the project's test/ directory to understand conventions. While this means the agent processes local code, this is inherent to its purpose as a development tool and presents minimal risk in a local context.
[Command Execution] (LOW): The skill recommends executing bin/rails test commands. These are standard developer operations and do not include arbitrary or dangerous command injection vectors.

rails-testing